Immobilized force sensor experiments

The next step was to demonstrate unfolding of hairpins in response to externally applied forces.

We began with a simple system of immobilizing DNA on a surface, which would then be subjected to forces. The advantage of this system is its minimum complexity, as we work with a single component. These experiments would also be useful in verification of the force range required for unfolding hairpins.

Plasma cleaned coverslips were silanized and coated with sulfo-SMCC to bind thiol modified DNA following a protocol described in detail in Chapter 6. A schematic of the coating protocol

2.6 Immobilized force sensor experiments 31

Figure 2.8: Single molecule controls of DNA sensor. Confocal scans of immobilized (A) HF strands and (B) HFQ strands show clear differences in the fluorescence intensity after quenching. (C) and (D) represent histograms of maximum photon counts per pixel from the identified 2D Gaussian single molecule patterns from (A) and (B), respectively.

is shown in Fig. 2.9A. The binding specificity of the coated coverslip was tested by spotting different DNA strands such as HF, HFQ, HFQC and washing thoroughly. F strand with biotin modification, instead of thiol, was also spotted as a negative control. Fluorescence microscopy (data not shown) revealed that the biotinylated strand was almost completely washed off, while the thiolated DNA remained attached to the surface. This verified the specificity of the coated coverslip for thiol modified DNA. For certain experiments the coverslip was then attached to a sticky-Slide I Luer (Ibidi GmbH, Martinsried, Germany) with a channel height of approximately 100 µm.

We confirmed immobilization and in situ hybridization of DNA strands with the following experiment. HF strands were spotted on coated coverslip and images were recorded in an epifluorescence microscope. Q strands were flushed into the channel and washed after 1 hr of incubation. Images of the same spot were recorded and the procedure was repeated af-ter flushing in control strands. Fluorescence intensities from images of HF, HFQ and HFQC

Figure 2.9: Immobilization of thiolated DNA on coated coverslips. (A) Plasma coated coverslips were silanized with APTES, followed by sulfo-SMCC treatment. Biotinylated HF, thiol ended HF, HFQ and HFQC strands were spotted on coverslip and enclosed in a 100µm channel. (B) Hybridization of immobilized DNA is demonstrated by spotting thiolated HF strand and sequentially adding Q and C strands. Images were recorded after 1 hr on addition of each type of strands. Intensities show quenching and recovery of fluorescence after flushing in Q and C strands, respectively. Error bars show standard error of mean.

strands are shown in Fig. 2.9B. Quenched HFQ strands showed reduced fluorescence intensity in comparison to HF strands. The addition of control strands (HFQC) resulted in an increase in the fluorescence intensity. This was consistent with our observations from bulk experiments in spectrophotometer. Thus, we could confirm that DNA was successfully immobilized and capable of hybridization with complementary strands in its immobilized state.

The next step was to unfold the immobilized hairpins using external force and record the corresponding changes in fluorescence. We tested various methods of force applications in parallel for the same; elaborated below. The reversible DNA sensor was used in all the following experiments, unless mentioned otherwise.

2.6.1 Magnetic beads experiments

A quick and easy method of exerting forces was to use a magnet to pull on magnetic beads attached to the DNA sensor. Streptavidin coated magnetic beads were obtained commercially as Dynabeads M-280 Streptavidin (Life Technologies GmbH, Darmstadt, Germany). Fig. 2.10A shows a schematic of the experimental set-up. HF strands were immobilized on coated coverslips and hybridized with biotinylated Q strands. The streptavidin coated magnetic beads were then attached to the immobilized DNA sensors. A small magnet was brought in vicinity of the coverslip, while simultaneously recording images. The presence of magnet resulted in visible movement of beads, except the few which remained strongly adhered to the substrate. However, the beads showed a strong autofluorescence at an excitation wavelength of 494 nm, which

2.6 Immobilized force sensor experiments 33

Figure 2.10: Magnetic beads to pull on immobilized DNA sensors. (A) Streptavidin coated magnetic beads were incubated with immobilized HFQ strands (Q strands with modified biotin ends). Beads were subjected to magnetic forces using a small board magnet. (B) Magnetic beads were autofluorescent at the excitation wavelength of the sensor, which probably masked any changes in the fluorescence intensity given by opening of the hairpins.

Figure 2.11: AFM experiments on immobilized DNA sensors. (A) An AFM cantilever tip was coated with neutravidin and interacted with HFQ (biotinylated Q) strands that were immobilized on a coverslip. (B) A representative force curve was measured in contact mode with an indentation force of 300 pN, waiting time of 1 s and retraction distance of 500 nm. The tip reacts non-specifically with the surface resulting in large forces during retraction; possibly ripping the immobilized DNA strands off the surface.

overlaps with the spectrum of fluorescent DNA, as shown in Fig. 2.10B. Thus, any fluorescence from opening of hairpins due to movement of beads would be masked and difficult to detect in presence of the autofluorescent beads.

2.6.2 AFM experiments

Another strategy was to apply forces on immobilized DNA sensor using atomic force microscopy (AFM). AFM has been commonly used to determine ligand-protein interaction forces by us-ing functionalized cantilever tips.^42–45 In our experiments, we used neutravidin functionalized cantilever tip to pull on biotinylated DNA strands, as described below.

HFQ complex was hybridized in a final concentration of 1 nM using F strand with thiol modification at end, hairpin and quencher strand with biotinylated end. The strands were immobilized on a sulfo-SMCC coated coverslip. The cantilever (Olympus Biolever RC150VB) was coated with neutravidin using a protocol described in Chapter 6. Schematic of experimental set-up is shown in Fig. 2.11A.

Experiments were performed on a MFP3D AFM set-up (Asylum Research, Santa Barbara, CA) combined with a TIRF microscope. The details of the experimental set-up are described elsewhere.⁴⁶The TIRF microscope was equipped with a CFI Apochromat TIRF 100x lens (Nikon GmbH, D¨usseldorf) and an Andor DL-658M-OEM (Andor Technology, Belfast, UK) camera.

Samples were excited with a 488 laser and images were recorded with an exposure time of 0.1 s. We operated AFM in contact mode with an approach force of 300 pN, waited for 1 s and retracted the cantilever for 500 nm at a velocity of 1 µm/s. A representative force curve obtained at these settings is shown in Fig. 2.11B. As seen from the force retraction curve, the cantilever tip adhered to the surface with non-specific interactions. Detachment of tip from the surface required a large force (approximately 500 pN) that probably also resulted in ripping off the immobilized DNA strands from the surface.

We also experienced problems in optimizing concentration of DNA strands for our exper-iments. As seen from bulk experiments, quenched DNA shows some basal level fluorescence.

High concentration of immobilized DNA resulted in a large background signal in the TIRF mi-croscope, making it difficult to detect any changes in fluorescence. Low concentration of DNA, on the other hand, resulted in reduced number of hairpins that could interact with the AFM tip.

2.6 Immobilized force sensor experiments 35

2.6.3 Flow experiments

Figure 2.12: Shear flow to unfold DNA hair-pins. Silica beads of 5µm diameter were coated with neutravidin and at-tached to biotin ends of Q strand on an immobilized HFQ complex. The DNA strands with attached beads were enclosed in a 100 µm high channel. A syringe pump was used to flow buffer into the channel at a constant flow rate.

Lastly, we used shear flow in a microfluidic chamber to apply forces on DNA hairpins.

HFQ strands; thiol modified F strands and Q strands with modified biotin ends were spot-ted on a coaspot-ted coverslip. Silica beads of 5 µm diameter were coated with neutravidin and then incubated with immobilized HFQ strands to allow interaction of streptavidin with biotinylated ends of Q strands. Inclu-sion of beads was necessary to ensure suffi-cient shear forces on the construct. This is because DNA is small in size, with a maxi-mum contour length of approximately 25 nm.

As the flow velocity close to the surface is neg-ligible, the hydrodynamic forces at surface of the chamber would be insufficient to open the hairpins. DNA strands, along with attached beads, were enclosed in a 100 µm channel with sticky-Slide I Luer (Ibidi GmbH, Martin-sried, Germany). A syringe pump (Harvard Apparatus, Holliston, MA, USA) was used to maintain a fixed flow rate. We tested a wide range of flow rates and yet did not observe an

increase in fluorescence intensity of the DNA sensors.

In this chapter, we discussed the need for novel 3D force sensor designs to explore force sensing and stress mapping in cells. DNA-based force sensor designs provide various advan-tages over protein-based ones; most importantly, flexibility in designs and possibility to detect a broad force range. We have, to this end, designed a simple DNA-based hairpin sensor incorpo-rating a FRET pair to give a visual output on application of force. We could also successfully demonstrate working of our sensor with control experiments in bulk, as well at single molecule level.

The force application experiments on immobilized DNA turned out to be more challenging than expected. AFM and magnetic bead experiments posed fundamental problems, making these methods unsuitable for our experiments. In shear flow experiments, estimated forces for unfolding the DNA hairpin are based on the assumption that the sensor is oriented perpendicular to the surface. Unfortunately, we have no information regarding orientation of the sensor. We speculate that horizontally oriented sensor that remains stuck in the plane of the coverslip might potentially hinder opening of hairpin on application of shear force.

Successfully unfolding immobilized DNA hairpins would have demonstrated working of the DNA sensor in a simple system. However, unprecedented situations added to complications,

demanding time consuming controls before further execution of these experiments. As we aim to measure forces in a 3D environment in cells, we redirected our focus to design an experimental system that better resembles cellular interior. In vitro composite networks of crosslinked cytoskeletal filaments have been shown to resemble living cells in their mechanical properties.⁴⁷Attaching DNA force sensors toin vitro cytoskeletal filament networks, which can be perturbed by external forces, would serve as an ideal system for force sensing experiments.

Moreover, crosslinking cytoskeletal filaments with DNA gives a model network with tunable crosslinker parameters.

Therefore, we generated in vitro networks of microtubules crosslinked with DNA sensor.

This system serves dual goals of providing an ideal experimental system to test force sensing as well as for studying mechanics of composite networks. The crosslinking and characterization of mechanics of these networks is described in detail in Chapter 3. The next section describes experiments monitoring response of DNA sensor in microtubule networks, after application of external shear forces.